Fault diagnosis of rolling bearings based on CFasterVit-TFAM and COS-UMAP model

PDF(2796 KB)

Journal of Vibration and Shock ›› 2025, Vol. 44 ›› Issue (10) : 287-300.

FAULT DIAGNOSIS ANALYSIS

Fault diagnosis of rolling bearings based on CFasterVit-TFAM and COS-UMAP model

QI Xiaoli, CUI Dehai*, WANG Zhiwen, ZHAO Fangxiang, WANG Zhaojun, MAO Junyi, YANG Wenhao

Author information +

History +

Abstract

Given the issues of imbalanced attention mechanisms, conservative pooling strategies, and the loss function's inability to comprehensively consider information from all classes leads to the learned features being relatively scattered in the FasterVit network, a rolling bearing fault diagnosis method based on the CFasterVit-TFAM and COS-UMAP models is proposed. The model consists of the FasterVit-TFAM network, the COS-UMAP dimensionality reduction algorithm, and the activation function CMSD-Softmax. Firstly, a new attention mechanism TFAM is proposed and combined with the FasterVit network to improve the balance and representation ability of information attention in the FasterVit network. Secondly, the COS-UMAP dimensionality reduction algorithm is used to replace the last pooling operation before the fully connected layer of the FasterVit network, effectively filtering and retaining important features in multidimensional data. Finally, replacing the cross-entropy loss function in the Softmax activation function with the mean standard deviation loss function allows for a more comprehensive learning of features and improves the model's generalization. The XJTU rolling bearing mixed fault experiment results show that the diagnostic accuracy of the TFAM attention mechanism is increased by 8% compared to other attention mechanisms, and the diagnostic accuracy of the COS-UMAP is increased by 15.8% compared to other dimensionality reduction algorithms. The diagnostic accuracy of the CMSD is increased by 0.5% compared to the cross entropy loss function. The proposed model achieves a recognition accuracy of 99.6% for fault samples, which is 1.4% higher than that of FasterVit and 7.8% higher than that of other network models. The simulation results of the rolling bearing dataset from Southeast University show that the proposed model achieves a recognition rate of 98.6% for fault samples, which is 2.2% higher than that of FasterVit. The average training time per round is reduced by 16.92 seconds, which is a maximum improvement of 12.2% compared to other network models, effectively improving the accuracy and generalization performance of the rolling bearing fault diagnosis model.

Key words

Fault diagnosis / Rolling bearings / FasterVit / Attention mechanism / Uniform Manifold Approximation and Projection / Class-distance Mean Standard Deviation loss

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

QI Xiaoli, CUI Dehai, WANG Zhiwen, ZHAO Fangxiang, WANG Zhaojun, MAO Junyi, YANG Wenhao. Fault diagnosis of rolling bearings based on CFasterVit-TFAM and COS-UMAP model[J]. Journal of Vibration and Shock, 2025, 44(10): 287-300

References

[1] 韩康,战洪飞,余军合,等.基于空洞卷积和增强型多尺度特征自适应融合的滚动轴承故障诊断[J].浙江大学学报(工学版),2024,58(06):1285-1295.
HAN Kang, ZHAN Hongfei, YU Junhe, et al. Fault diagnosis of rolling bearings based on dilated convolution and enhanced multi-scale feature adaptive fusion [J]. Journal of Zhejiang University (Engineering Edition), 2024,58 (06): 1285-1295.
[2] 李军宁,罗文广,陈武阁.面向振动信号的滚动轴承故障诊断算法综述[J].西安工业大学学报,2022,42(02):105-122.
LI Junning, LUO Wenguang, CHEN Wuge. Overview of fault diagnosis algorithms for rolling bearings based on vibration signals [J]. Journal of Xi'an University of Technology, 2022, 42 (02): 105-122.
[3] Sahu P K , Rai R N .LSTM-based deep learning approach for remaining useful life prediction of rolling bearing using proposed C-MMPE feature[J].Journal of Mechanical Science and Technology, 2024, 38(5):2197-2209.
[4] CHEN X, SHU G, ZHANG K, et al. A fault characteristics extraction method for rolling bearing with variable rotational speed using adaptive time-varying comb filtering and order tracking[J]. Journal of Mechanical Science and Technology,2022, 36(3): 1171-1182.
[5] 王鹏,李丹青,王恒.基于改进交替迁移学习的滚动轴承故障诊断算法[J].振动与冲击,2024,43(05):239-249.
WANG Peng, LI Danqing, WANG Heng. Fault diagnosis algorithm for rolling bearings based on improved alternating transfer learning [J]. Vibration and Shock, 2024,43 (05): 239-249.
[6] 张英杰,张彩华,陆碧良,等.基于类别域自适应的滚动轴承故障诊断[J].振动与冲击,2023,42(24):117-126.
ZHANG Yingjie, ZHANG Caihua, LU Biliang, et al. Fault diagnosis of rolling bearings based on category domain adaptation [J]. Vibration and Impact, 2023,42 (24): 117-126.
[7] 火久元,李宇峰,常琛,等.基于混合裁剪失衡数据增强与SwinNet网络的滚动轴承故障诊断[J].振动与冲击,2024,43(06):64-74.
HUO Jiuyuan, LI Yufeng, CHANG Chen, et al. Fault diagnosis of rolling bearings based on mixed cropping imbalanced data augmentation and SwinNet network [J]. Vibration and Shock, 2024,43 (06): 64-74.
[8] 康玉祥,陈果,盛嘉玖,等.低转速航空发动机滚动轴承故障深度异常检测方法[J].振动与冲击,2024,43(07):186-195.
KANG Yuxiang, CHEN Guo, SHENG Jiajiu, et al. Detection method for deep abnormal faults in low speed aircraft engine rolling bearings [J]. Vibration and Impact, 2024,43 (07): 186-195.
[9] 邓飞跃,陈哲,郝如江,等.基于MsTCN-Transformer模型的轴承剩余使用寿命预测研究[J].振动与冲击,2024,43(04):279-287.
DENG Feiyue, CHEN Zhe, HAO Rujiang, et al. Prediction of Remaining Service Life of Bearings Based on the MsTCN Transformer Model [J]. Vibration and Impact, 2024,43 (04): 279-287.
[10] 周臻,顾子渊,曲小波,等.城市多模式交通大模型MT-GPT:点线面的分层技术与应用场景[J].中国公路学报,2024,37(02):253-274.
ZHOU Zhen, GU Ziyuan, QU Xiaobo, et al. Multi modal Urban Traffic Model MT-GPT: Layered Technology and Application Scenarios of Points, Lines, and Planes [J]. Chinese Journal of Highways, 2024,37 (02): 253-274.
[11] 张帆,姚德臣,姚圣卓,等.基于Transformer-LSTM网络的轴承寿命预测[J].振动与冲击,2024,43(06):320-328.
ZHANG Fan, YAO Dechen, YAO Shengzhuo, et al. Bearing Life Prediction Based on Transformer LSTM Network [J]. Vibration and Impact, 2024,43 (06): 320-328.
[12] Ghahremani M, Khateri M, Jian B, et al. H-ViT: A Hierarchical Vision Transformer for Deformable Image Registration[C]//P-roceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2024: 11513-11523.
[13] 韩争杰,牛荣军,马子魁,等.基于注意力机制改进残差神经网络的轴承故障诊断方法[J].振动与冲击,2023,42(16):82-91.
HAN Zhengjie, NIU Rongjun, MA Zikui, et al. A bearing fault diagnosis method based on attention mechanism improved residual neural network [J]. Vibration and Shock, 2023,42 (16): 82-91.
[14] 姚齐水,别帅帅,余江鸿,等.一种结合改进InceptionV2模块和CBAM的轴承故障诊断方法[J].振动工程学报,2022,35(04):949-957.
YAO Qishui, BIE Shuaishuai, YU Jianghong, et al. A bearing fault diagnosis method combining improved InceptionV2 module and CBAM [J]. Journal of Vibration Engineering, 2022, 35 (04): 949-957.
[15] 王敏,邓艾东,马天霆,等.基于双注意力机制的MSCN-BiGRU的滚动轴承故障诊断方法[J].振动与冲击,2024,43(06):84-92+103.
WANG Min, DENG Aidong, MA Tianting, et al. Fault diagnosis method for rolling bearings based on dual attention mechanism MSCN BiGRU [J]. Vibration and Shock, 2024,43 (06): 84-92+103.
[16] Yan W ,Jie L ,Xiaoguang G , et al. Multi-scale attention mechanism residual neural network for fault diagnosis of rolling bearings [J]. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 2022, 236 (20): 10615-10629.
[17] 冯立伟,孙立文,顾欢,等.基于增量式等距映射同双重局部密度方法的工业过程故障检测[J].上海交通大学学报,2024,58(04):525-533.
FENG Liwei, SUN Liwen, GU Huan, et al. Industrial process fault detection based on incremental isometric mapping and dual local density method [J]. Journal of Shanghai Jiao Tong University, 2024,58 (04): 525-533.
[18] 姜战伟,郑近德,潘海洋,等.基于多尺度时不可逆与t-SNE流形学习的滚动轴承故障诊断[J].振动与冲击,2017,36(17):61-68+84.
JIANG Zhanwei, ZHENG Jinde, PAN Haiyang, et al. Fault diagnosis of rolling bearings based on multi-scale irreversible and t-SNE manifold learning [J]. Vibration and Shock, 2017, 36 (17): 61-68+84.
[19] 施莹,庄哲,林建辉.基于卷积稀疏表示及等距映射的轴承故障诊断[J].振动.测试与诊断,2019,39(05):1081-1088+1138.
SHI Ying, ZHUANG Zhe, LIN Jianhui. Bearing Fault Diagnosis Based on Convolutional Sparse Representation and Isometric Mapping [J]. Vibration. Testing and Diagnosis, 2019,39 (05): 1081-1088+1138.
[20] 李建斌,武颖莹,李鹏宇,等.基于局部线性嵌入和支持向量机回归的TBM施工参数预测[J].浙江大学学报(工学版),2021,55(08):1426-1435.
LI Jianbin, WU Yingying, LI Pengyu, et al. Prediction of TBM construction parameters based on local linear embedding and support vector machine regression [J]. Journal of Zhejiang University (Engineering Edition), 2021,55 (08): 1426-1435.
[21] 刘璐,邱明,李军星,等.基于LTSA融合降维法的滚动轴承可靠性评估方法[J].航空动力学报,2021,36(02):413-420.
LIU Lu, QIU Ming, LI Junxing, et al. Reliability evaluation method for rolling bearings based on LTSA fusion dimension reduction method [J]. Journal of Aerodynamics, 2021,36 (02): 413-420.
[22] Mcinnes L , Healy J .UMAP: Uniform Manifold Approxima-tion and Projection for Dimension Reduction[J].The Journal of Open Source Software, 2018, 3(29):1-63.
[23] 马文源,袁蜀翔,刘宁,等.基于改进卷积神经网络的齿轮箱故障诊断方法[J].自动化与仪器仪表,2023,(09):46-50.
MA Wenyuan, YUAN Shuxiang, LIU Ning, et al. Gearbox Fault Diagnosis Method Based on Improved Convolutional Neural Network [J]. Automation and Instrumentation, 2023, (09): 46-50.
[24] 齐咏生,巩育瑞,高胜利,等.基于中心损失-改进卷积自编码器的滚动轴承半监督故障诊断[J].振动与冲击,2023,42(07):301-311.
Qi Yongsheng, Gong Yurui, Gao Shengli, et al. Semi supervised fault diagnosis of rolling bearings based on center loss improved convolutional autoencoder [J]. Vibration and Shock, 2023, 42 (07): 301-311.
[25] 张康智.基于联合LLE和SSR的滚动轴承故障诊断方法[J].航空动力学报,2024,39(07):41-49.
Zhang Kangzhi. Fault diagnosis method for rolling bearings based on joint LLE and SSR [J]. Journal of Aerodynamics, 2024, 39 (07): 41-49.
[26] 徐鹏,皋军,邵星.基于AMCNN-BiGRU的滚动轴承故障诊断方法研究[J].振动与冲击,2023,42(18):71-80.
XU Peng, GAO Jun, SHAO Xing. Research on Fault Diagnosis Method for Rolling Bearings Based on AMCNN BiGRU [J]. Vibration and Impact, 2023,42 (18): 71-80.
[27] Gunasekaran P, Alamdari M M, Goudarzi H V P. Ultrasonic based defect detection in steel-reinforced laminated timber structural elements using Uniform Manifold Approximation and Projection (UMAP)[J]. Automation in Construction, 2024, 160: 105296.
[28] You P, Yang R. A Semi-supervised Fault Diagnosis Method Based on Deep Adaptation Autoencoder and Manifold Learning for Rolling Bearings[C]//2022 27th International Conference on Automation and Computing (ICAC). IEEE, 2022: 1-6.
[29] 高玉才,付忠广,谢玉存,等.基于教师-学生网络的半监督故障诊断模型[J].振动与冲击,2024,43(04):150-157.
GAO Yucai, FU Zhongguang, XIE Yucun, et al. A semi supervised fault diagnosis model based on teacher student network [J]. Vibration and Shock, 2024,43 (04): 150-157.
[30] Yao S , Zhang Z , Han B ,et al.Fast nonlinear cross-sparse filtering for rolling bearings compound fault diagnosis[J].Me-asurement Science & Technology, 2024(3):35-50.